ASTRONOMY: H. N. RUSSELL 
75 
of the terrestrial group. For Jupiter and the remoter planets, however, 
estimates can be made which are not Hkely to be much more than 15% in 
error. This factor depends mainly upon the texture of the planet's 
surface, being high if this is smooth, and low if it is rough and covered 
with irregularities whose shadows darken considerable areas at phases 
remote from the full. 
3. The relative brightness of the sun and stars, as seen from the 
earth, has been determined with surprising accuracy, — the results of 
several observers, by radically different methods, being in excellent 
agreement. The sun's stellar magnitude, on the Harvard scale, accord- 
ing to the mean of the observations of Zollner, Ch. Fabry, Ceraski and 
W. H. Pickering, is —26.72 =t= 0.04, — which is equivalent to saying that 
the sun appears to be 123 thousand million times as bright as a standard 
first magnitude star. The photographic magnitude of the sun, accord- 
ing to King and Birck, is —25.93, and its color index +0.79, agreeing 
very closely with the average for stars of similar spectrum (Class G) . 
4. The law of variation of the moon's brightness with phase is very 
well determined by the observations of J. Herschel, Bond, Zolhier, W. H. 
Pickering, King, Stebbins and Brown, and Wislicenus, — the results of 
all seven agreeing satisfactorily with a mean curve. The full moon is 
8.7 times brighter than the first quarter, and 10.0 times brighter than the 
last quarter. The remarkable falHng off in brightness between the full 
and half moon shows that the lunar surface must be very rough, as was 
first pointed out by Zollner. The difference between the waxing and wan- 
ing moon arises from the greater extent of the dark maria on the eastern 
half of the visible disk (as Stebbins has shown). 
5. The results of different observers for the brightness of the mean 
full moon, compared with the sun or stars, are discordant. J. Herschel's 
observations^ have been reduced anew, and an error which had crept 
into the earher reduction corrected. The weighted mean of those 
determinations which are not obviously affected by systematic error 
makes the visual magnitude —12.55, and the ratio of sunlight to mean 
full moonlight 465,000, — with an uncertainty of fully 10%. The photo- 
graphic magnitude, —11.37, has been well determined by King. It 
shows that moonlight is redder than sunlight, — in agreement with the 
spectro-pho tome trie measures of Wilsing and Scheiner. 
6. Miiller's data for the major planets^ have been adopted, with a 
correction of —0.06 mag. to reduce to the Harvard scale, except for 
Uranus and Neptune; Pickering's magnitudes for the asteroids,^ and 
Guthnick's for the satellites of Jupiter and Saturn.^ The color indices of 
Venus, Mars, Jupiter and Saturn have been derived from comparison of 
Miiller's visual and King's photographic observations.® 
